Epidemiological aspects and spatial distribution of visceral leishmaniasis in Governador Valadares, Brazil, between 2008 and 2012

Abstract INTRODUCTION: Visceral leishmaniasis (VL) is an important parasitic disease. We evaluated the epidemiological aspects and spatial distribution of visceral leishmaniasis in Governador Valadares, Brazil. METHODS: All cases of VL, registered by the municipal health department, were analyzed and georeferenced. RESULTS: The human mortality rate was 15% and canine seroprevalence rate was 29.0%. Higher numbers of canine VL cases correlated with higher incidence of human cases. CONCLUSIONS: The high rate of canine seroprevalence, resurgence of the human disease, and correlation between canine and human VL reinforces the role of the dog in disease transmission within the municipality.

Visceral leishmaniasis (VL) is a neglected and endemic parasitic disease reported in 65 countries. In 2017, 94% of the new cases occurred in seven countries: Ethiopia, India, Kenya, Somalia, South Sudan, Sudan and Brazil 1 . In Brazil, the etiological agent of VL is Leishmania infantum, which is widely distributed in the wild and domestic environments 2 . The main vector is the Lutzomyia longipalpis sandfly, and dogs (Canis familiaris) are the main urban reservoir in domestic and peridomicile environments, contributing to the life cycle of the disease in urban areas 3 .
Dogs are responsible for endemic and epidemic VL in large urban areas 4 . The canine enzootic disease preceded the occurrence of human cases, and the infection has been more prevalent in dogs than in humans 5 .
The first human cases of VL in Vale do Rio Doce, Minas Gerais (MG) were reported in 1966, in a predominantly rural area 6 . Since then, no further cases were reported and the disease was considered to be under control. In 2008, the disease re-emerged in Governador Valadares, considered an endemic area due to the active transmission of VL 7 . To date, no epidemiological studies of leishmaniasis using spatial analysis have been performed. Spatial analysis would enable studying the superposition of human and canine VL cases, and this may aid the planning of disease control measures. Therefore, the present study evaluated the spatial distribution and epidemiological aspects of VL in the municipality of Governador Valadares, MG, between 2008 and 2012. Governador Valadares (18º 51' 2''S, 41º 56' 53''W) is located in the eastern region of the state of MG, in the mesoregion of Vale do Rio Doce. In 2010, the demographic census estimated the population to be 263,689 8 .
This study included all individuals diagnosed with VL, confirmed by the clinic-laboratory criteria, registered in the urban area by the epidemiology administration (GEPI) of the Department of Health Surveillance, Municipal Health Secretariat of the city of Governador Valadares, from 2008 to 2012. The serology techniques included immunochromatographic rapid test and indirect immunofluorescence reaction (RIFI) performed at the serology laboratory.
This study included all dogs diagnosed with VL, confirmed by the laboratory criteria, registered in urban areas by the local Zoonosis Control Division (ZCD). The serology techniques included enzyme-linked immunosorbent assays (ELISAs) and RIFI performed at the serology laboratory.
The Brazilian Ministry of Health recommends the demarcation of VL transmission areas in each municipality. These areas are stratified according to the average number of cases reported in the last three years. Municipalities with less than an average of 2.4 cases are classified as areas of sporadic transmission. Municipalities reporting between 2.4 and less than 4.4 cases and above or equal to 4.4 cases are considered areas of moderate and active transmission, respectively 9 .
All data necessary for the sectorization of the municipality and drawing of maps were made available by the ZCD and GEPI of the Department of Health Surveillance, Municipal Health Secretariat of the city of Governador Valadares. The municipality of Governador Valadares comprises 150 neighborhoods. Based on their proximity and similarities in socioeconomic and environmental characteristics, the neighborhoods were grouped into nine sectors (A to I).
The sites of human and canine VL cases were organized according to streets, buildings, and neighborhoods, and linked to the municipal land registry office (LRO) through a unique identification number for each plot of land. These numbers were generated in the municipal register database using the NetTerm program. CTM facilitated the precise location of the plots to be geocoded and the production of elaborate thematic maps.
To map human and canine VL cases, the codes for buildings and public places were organized in Excel spreadsheets and georeferenced using the ArcGIS software (ESRI, Redlands, CA, USA) developed for visualization of Geographic Information Systems (GIS) data.
Statistical analyses were performed using Excel spreadsheets and MedCalc statistical software (MedCalc Software bvba, Ostend, Belgium) (https://www.medcalc.org/calc/). The Pearson linear correlation coefficient was used to measure the following correlations: a) canine population and human VL cases; b) canine population and canine VL cases; c) canine and human VL cases. The municipality of Governador Valadares consists of 150 neighborhoods, and these were grouped into nine sectors (A to I) based on their proximity and socioeconomic and environmental characteristics (Figure 1).
GEPI reported 115 cases of human VL between 2008 and 2012 ( Table 1). The first case occurred in 2008. There were 14,  Correlation analyses showed that a correlation exists between the canine population and canine VL cases (0.9918) and also between canine and human VL cases (0.973), significant at 0.1%.
The number of dogs in the municipality did not correlate with a higher incidence of human VL cases. However, the highest number of canine VL cases correlated with a higher incidence of human VL cases. The reported canine seroprevalence was 5.8% between 2008 and 2009 in the municipality of Montes Claros 10 . This rate was lower than that reported in Governador Valadares, which was 25.5% during the same period (data not shown). In 2015, the seroprevalence was reportedly 14.8% in the municipality of Ipatinga, about 105 km from Governador Valadares 12  A limitation of this study is the absence of a statistical analysis demonstrating the sector-wise correlation between human and canine cases. The sectorization of the municipality of Governador Valadares in the present study allowed a better visualization of the disease distribution. All sectors had a high prevalence of canine VL cases. However, in sectors with a higher incidence of human cases (A, B and D), rates of canine seroprevalence were even higher. Sectorization by neighborhoods has also been conducted in other municipalities, such as Janaúba, MG, which is considered an area of intense VL transmission 15 .
Therefore, the results of this study indicated that dogs play an important role in the transmission of human VL. Furthermore, a reassessment of the disease control strategies, using the epidemiological surveillance of the municipality, would be of immense interest.

ACKNOWLEDGMENTS
GEPI and ZCD -Department of Health Surveillance, Municipal Health Secretariat of the city of Governador Valadares.